Professional Documents
Culture Documents
Abstract—Detection of helmeted and non-helmeted motorcy- and violation of that attracts a hefty fine. Currently, all major
clist is mandatory now-a-days in order to ensure the safety of cities employ surveillance system on roads to identify violators
riders on the road. However, due to many constraints such as poor which needs a system that can automatically detect whether a
video quality, occlusion, illumination, and other varying factors it
becomes very difficult to detect them accurately. In this paper, we motorcyclist is wearing a helmet or not. This paper aims to an
introduce an approach for automatic detection of helmeted and automatic helmet and non-helmet detection of motorcyclists.
non-helmeted motorcyclist using convolutional neural network
(CNN). During the past several years, the advancements in deep II. R ELATED W ORK
learning models have drastically improved the performance of Since past several years, many researchers have solved
object detection. One such model is YOLOv2 [1] which combines
both classification and object detection in a single architecture. the problem of helmet detection. Wen et al. [4] proposed
Here, we use YOLOv2 at two different stages one after another a circular arc detection method based on Hough transform
in order to improve the helmet detection accuracy. At the first for helmet detection. Chiu et al. [5] use a Canny edge
stage, YOLOv2 model is used to detect different objects in the detector [6] to detect helmets. Similarly, authors in [7] use
test image. Since this model is trained on COCO dataset, it a background subtraction to detect moving vehicles. Here,
can detect all classes of the COCO dataset. In the proposed
approach, we use detection of person class instead of motorcycle authors calculate a image based statistical information by
in order to increase the accuracy of helmet detection in the input isolating the approximate region of the head of the riders.
image. The cropped images of detected persons are used as input In order to extract features, histogram of gradient (HOG)
to second YOLOv2 stage which was trained on our dataset of descriptors is used from the isolated regions and then results
helmeted images. The non-helmeted images are processed further are passed to a support vector machine (SVM) classifier for
to extract license plate by using OpenALPR. In the proposed
approach, we use two different datasets i.e., COCO and helmet classification purpose. Furthermore, authors in [8] proposed
datasets. We tested the potential of our approach on different a model for vehicle detection, tracking and classification
helmeted and non-helmeted images. Experimental results show from roadside CCTV cameras. In order to make their model
that the proposed method performs better when compared to invariant against illumination changes and camera vibrations,
other existing approaches with 94.70% helmet detection accuracy. they use a new background Gaussian mixture model. Silva
Index Terms—helmet detection, YOLOv2, license plate extrac-
tion, COCO et al. in [9] propose a hybrid descriptor model based on
geometric shape and texture features for automatic detection of
I. I NTRODUCTION motorcyclists without helmet. They use Hough transform with
SVM to detect the head of the motorcyclist. Additionally, they
Motorcycles being one of the most convenient modes of extend their work in [10] by multi-layer perception model for
transport has led to increasing use and thus accounts for the classification. In [11], Waranusat et al. [11] propose a system
highest share of total road accidents. As per one survey carried to detect moving objects with k-NN classifier over the head
out in the year of 2014, 30% of all road accidents related to region of motorcyclist to classify helmet. Similarly, authors
deaths were of riders on two-wheelers [2]. Also, as per the in [12] use background subtraction and object segmentation
record of Chennai, India, between the period of January 1, with k-NN classifier to detect bike-riders. All these methods
2013 to June 28, 2015, total 1,453 two-wheeler riders who died which are employed with engineered features have limitations
in road accidents were not wearing helmet [3]. Also, according in accurate detection of helmets.
to the World Health Organization (WHO), wearing a helmet Alex et al. [13] introduced a convolutional neural network
can reduce the risk of severe injury by 72% and the risk of (CNN) based method for object classification and detection.
death by 39% [2]. In India, road safety has been a neglected Since then CNNs became the standard model for classification
area. Due to this wearing a helmet has been made mandatory as it outperformed all the traditional models such as HOG
978-1-5386-1842-4/17/$31.00 2017
c IEEE [14], scale-invariant feature transform (SIFT) [15] and local
binary patterns (LBP) [16]. Recently, A. Hirota et al. [17] use
a CNN for classification of helmeted and non-helmeted riders.
Although they use CNN, their helmet detection accuracy is
poor with limitations of helmet color and multiple riders on
a single motorcyclist. Literature shows that the CNN can
learn engineered features from raw data effectively and it
outperforms over it’s hand-crafted counterparts. Therefore,
we use CNN to solve helmet detection problem. The overall
contributions of our paper are as follows:
• Use of person detection instead of motorcyclist in order to
increase the helmet detection accuracy in the test image.
• Training a model using our database of helmeted images.
• Extraction of license number plate for non-helmeted
motorcyclist.
same is depicted in Fig. 4. We tested the model at various Fig. 4: Quantitative measures during training phase.
checkpoints to obtain the results. One can observe in Fig. 4
that starting from 500 iterations, recall and accuracy increase
rapidly. However, these values remain steady after 800 itera- reduced exponentially and finally it becomes stagnant, with a
tions. The graph of precision begins with high value since most very little but significant change, after 500 iterations. We stop
of the predictions are false in the initial phase of training then our training around 2500 iterations where the error was not
it decreases slightly and it remains constant for the most part changing much. Here, we use a learning rate of 0.001.
of the graph. The final values of those quantitative measures Additionally, the graph of average of intersection-over-union
are displayed in Table I. One can note that helmet detection (IOU) against number of iterations is depicted in Fig. 6. Here,
accuracy using proposed method is 94.70% which is better we use average of IOU over every 150 iterations, and plot the
when compared to other previous state of the art approaches same against the sampled units which is obtained by dividing
[9]–[11]. Silva et al. [9] obtain helmet detection accuracy of the number of iterations by 150. Here, the average IOU
94.23% using SVM classifier. Authors in [10] use multi-layer increases rapidly in the initial stage of training. Then it starts
perceptron model for classification. They obtain the helmet to fluctuate by 2% to 4%, while still generally increasing, and
detection accuracy of 91.37% on their own database of 255 finally reaches to around 80%, after which the error and IOU
images. Furthermore, Waranusat et al. [11] obtain accuracy remains constant. Finally, Fig. 7 shows the results of helmet
of 74% by using k-NN classifier over the head region of detection at various phases of training. Fig. 7(a) displays the
motorcyclist. result obtained when 150 iterations of training for helmet
In Fig. 5, we display the graph of average error during detection are completed. Here, no helmet is detected. When
training iterations i.e., number of batches completed. Here, iteration reach to 300 iterations, some helmets are detected as
one can see that at the beginning of the training, the average displayed in Fig. 7(b). Interestingly, proposed approach detects
error started very high, i.e., approximately 160, and then it all helmets properly when 950 iterations are over as displayed
V. C ONCLUSION
We propose an approach for automatic helmet detection
using CNN. In order to increase the helmet detection accuracy
we use two stage of YOLOv2 models. First YOLOv2 model
ensures the person detection which is trained on COCO
dataset. This step decreases the number of helmets being
undetected. The cropped images of detected person are used as
input to the second YOLOv2 model which was trained on our
helmeted image dataset. The proposed method has been tested
on different helmeted image scenarios. Also, we evaluated the
quantitative measures on the test images and same is compared
with other state of the art methods. Experimental results and
quantitative measures on different scenarios show that our
Fig. 5: Average error Vs number of iterations. approach is robust and reliable with a high helmet detection
accuracy.
ACKNOWLEDGEMENT
Authors are gratefully acknowledge the support of NVIDIA
Corporation with the donation of the Titan Xp GPU used for
this research.
R EFERENCES
[1] J. Redmon and A. Farhadi, “YOLO9000: better, faster,
stronger,” CoRR, vol. abs/1612.08242, 2016. [Online]. Available:
http://arxiv.org/abs/1612.08242
[2] [Online]. Available: https://scroll.in/article/757365/road-accidents-kill-
382-in-india-every-day-1682-times-more-than-terrorism
[3] [Online]. Available: http://timesofindia.indiatimes.com/city/chennai/98-
6-of-bikers-who-died-didnt-wear-a-helmet/articleshow/47904790.cms
[4] C.-Y. Wen, S.-H. Chiu, J.-J. Liaw, and C.-P. Lu, “The safety helmet
detection for atm’s surveillance system via the modified hough trans-
form,” in IEEE 37th Annual 2003 International Carnahan Conference
onSecurity Technology, 2003. Proceedings., Oct 2003, pp. 364–369.
Fig. 6: Average IOU Vs number of iterations. [5] C.-C. Chiu, C.-Y. Wang, M.-Y. Ku, and Y. B. Lu, “Real time recogniton
and tracking system of multiple vehicles,” in 2006 IEEE Intelligent
Vehicles Symposium, 2006, pp. 478–483.
[6] J. Canny, “A computational approach to edge detection,” IEEE Transac-
tions on Pattern Analysis and Machine Intelligence, vol. PAMI-8, no. 6,
pp. 679–698, Nov 1986.
[7] J. Chiverton, “Helmet presence classification with motorcycle detection
and tracking,” IET Intelligent Transport Systems, vol. 6, no. 3, pp. 259–
269, September 2012.
[8] Z. Chen, T. J. Ellis, and S. A. Velastin, “Vehicle detection, tracking and
(a) (b) classification in urban traffic,” 2012 15th International IEEE Conference
on Intelligent Transportation Systems, pp. 951–956, 2012.
[9] R. Silva, K. Aires, T. Santos, K. Abdala, R. Veras, and A. Soares,
“Automatic detection of motorcyclists without helmet,” in 2013 XXXIX
Latin American Computing Conference (CLEI), Oct 2013, pp. 1–7.
[10] R. R. V. e. Silva, K. R. T. Aires, and R. d. M. S. Veras, “Helmet detection
on motorcyclists using image descriptors and classifiers,” in 2014 27th
SIBGRAPI Conference on Graphics, Patterns and Images, Aug 2014,
(c) (d) pp. 141–148.
[11] R. Waranusast, N. Bundon, V. Timtong, C. Tangnoi, and P. Pat-
Fig. 7: Detection of multiple helmets for different iterations: tanathaburt, “Machine vision techniques for motorcycle safety helmet
(a) 150 iterations (b) 300 iterations (c) 950 iterations (d) 2500 detection,” in 2013 28th International Conference on Image and Vision
Computing New Zealand (IVCNZ 2013), Nov 2013, pp. 35–40.
iterations. [12] K. Dahiya, D. Singh, and C. K. Mohan, “Automatic detection of bike-
riders without helmet using surveillance videos in real-time,” in 2016
International Joint Conference on Neural Networks (IJCNN), July 2016,
pp. 3046–3051.
in Fig. 7(c). However, the objects which are not helmet are also [13] A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet classification
detected as class ‘helmet’. Finally, Fig. 7(d) depicts the helmet with deep convolutional neural networks,” in Advances in Neural
detection results obtained when 2500 iterations are over which Information Processing Systems 25, F. Pereira, C. J. C. Burges,
L. Bottou, and K. Q. Weinberger, Eds. Curran Associates, Inc., 2012,
shows that the final weights detect all the helmets correctly pp. 1097–1105. [Online]. Available: http://papers.nips.cc/paper/4824-
without any false positive. imagenet-classification-with-deep-convolutional-neural-networks.pdf
[14] N. Dalal and B. Triggs, “Histograms of oriented gradients for human
detection,” in 2005 IEEE Computer Society Conference on Computer
Vision and Pattern Recognition (CVPR’05), vol. 1, June 2005, pp. 886–
893 vol. 1.
[15] D. G. Lowe, “Distinctive image features from scale-invariant keypoints,”
Int. J. Comput. Vision, vol. 60, no. 2, pp. 91–110, Nov. 2004. [Online].
Available: http://dx.doi.org/10.1023/B:VISI.0000029664.99615.94
[16] T. Ahonen, A. Hadid, and M. Pietikäinen, Face Recognition with Local
Binary Patterns. Springer Berlin Heidelberg, 2004, pp. 469–481.
[17] A. Hirota, N. H. Tiep, L. Van Khanh, and N. Oka, Classifying Hel-
meted and Non-helmeted Motorcyclists. Cham: Springer International
Publishing, 2017, pp. 81–86.
[18] T. Lin, M. Maire, S. J. Belongie, L. D. Bourdev, R. B. Girshick,
J. Hays, P. Perona, D. Ramanan, P. Dollár, and C. L. Zitnick, “Microsoft
COCO: common objects in context,” CoRR, vol. abs/1405.0312, 2014.
[Online]. Available: http://arxiv.org/abs/1405.0312
[19] GitHub. (2015) Openalpr/openalpr”. [Online]. Available:
https://github.com/openalpr/openalpr
[20] J. Redmon, S. K. Divvala, R. B. Girshick, and
A. Farhadi, “You only look once: Unified, real-time object
detection,” CoRR, vol. abs/1506.02640, 2015. [Online]. Available:
http://arxiv.org/abs/1506.02640
[21] J. Redmon, “Darknet: Open source neural networks in c,”
http://pjreddie.com/darknet/, 2013–2016.
[22] J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, “Imagenet:
A large-scale hierarchical image database,” in Computer Vision and
Pattern Recognition, 2009. CVPR 2009. IEEE Conference on. IEEE,
2009, pp. 248–255.